Terminal-Agent Documentation

Terminal-Agent is a framework to run LLM workflow in terminal environment, developed by Sea AI Lab.

Get Started

• Quickstart
• Autopilot CLI
• Run With Sandbox
• Tool Use for Agents

Basic Usage

• Web Application
• Annotation
• Evaluation
• Choose LLM Service

Task Reference

• Task Overview
• LoCoMo Task Design
• MMLU-Pro Task Design
• SWE-Bench Task Design
• Terminal-Bench Task Design

Developer Reference

• Design Philosophy
• Workflow
• Hooks
• Nodes

Developer Guideline

• Contribution Guide
• Code Standards